Tracking Fish Behavior in the Tailrace of a Hydropower Plant

While conditions in the tailraces of dams make studying fish telemetries difficult, new technologies employed at Brazil’s 66 MW Tres Marias project have given researchers a better insight into fish behavior and movement.

By Fabio Suzuki and Paulo Santos Pompeu

Dams can interrupt fish reproductive cycles by fragmenting habitat and migratory pathways. In particular, the tailraces of dams that impound water for hydropower facilities present significant challenges for migratory fish attempting to move upstream to spawn. The hydrodynamic conditions of the tailrace – including high turbulence and temperature variations – attract a large number of migratory fish to the tailrace.

Fish potentially can swim into the draft tubes during operation of the turbines, becoming trapped and exposed to low dissolved oxygen, which can lead to mortality. Other dangerous exposure can include the rapid decompression and mechanical shock experienced during the system’s starts and stops.

Detailed monitoring of fish movement and migration patterns in tailraces can provide information for reducing fish mortality in turbines and at other key points in tailraces.

However, little is known about upstream migrating fish in the tailraces of dams because the majority of fish telemetry technologies have been unable to detect fish with precision in the acoustically noisy tailraces of dams. The study described in this article sought to determine if new tracking technology could be used to track fish movement and better understand their behavior. The end goal would be for hydro project owners to gain insights about how to manage operations of hydropower facilities to minimize fish kill in the turbines.


The study brought together a number of parties, including the Federal University of Lavras (UFLA) Department of Biology; Brazilian power generator Companhia Energetica de Minas Gerais (Cemig); and fisheries biologists from Hydroacoustic Technology Inc. based in Seattle, Wash., USA.

The group set out to evaluate specific numbers of two migratory fish species that are abundant, important species for commercial fisheries and the most common species affected by turbine operation in Tres Marias Dam. For this study, 50 argenteus (Prochilodus argenteus) and 40 maculatus (Pimelodus maculatus) were equipped with HTI-provided acoustic tags that were capable of being tracked three-dimensionally in the acoustically-challenging aquatic environment.

The number of tagged fish was chosen for achieving a reasonable tracking data for fish movement analysis, taking into account the funding limitation and the possibility of tag loss and mortality rate by capture-tagging procedure. The tag weights and tag lifespans were ideal for the requirements of this study, and the 3D tracking technique used permitted numerous tags to be tracked simultaneously.

Researchers inserted HTI acoustic tags into 90 fish, allowing for tracking in three dimensions.
Researchers inserted HTI acoustic tags into 90 fish, allowing for tracking in three dimensions.

This evaluation was conducted in the tailrace of the 66 MW Tres Marias plant, located on the Sao Francisco River in southeastern Brazil. The project’s dam is a 2,700 m long and 75 m tall, with a mean flow of 150 m3/sec per turbine. The draft tube for each turbine is about 15 meters long, while the tailrace is about 155 meters long and 182 meters wide, giving it an area of about 19,870 m2. The dam is not equipped with fish passage facilities.

From Oct. 31 to Nov. 4 in 2011, fish were captured in the tailrace and up to 3 km downstream in the Sao Francisco River using cast nets. Acoustic tags were then surgically inserted into the fishes’ body cavities in a controlled environment.

The two tagged species were then evaluated for daily variations in spatial distributions, depth preferences and the influence of turbine discharge on the spatial distribution of fish in the tailrace from Oct. 31, 2011, through Feb. 16, 2012.

All the 11 hydrophones used to capture data were geo-referenced and connected to an HTI Model 20 Acoustic Tag Tracking System Data Logger. The hydrophones position was based on data gathered from previous tests conducted in tailrace, which indicated a detection range of 100 m. The maximum distance between adjacent hydrophones was about 85 meters. Additionally, the hydrophones were installed in different depths, with five close to the bottom and six close to surface. This system operated at 307 kHz and it continuously received and stored all tag transmission pulses for each of the hydrophones deployed. The telemetry system operated 24 hours per day and was able to calculate the dimensions of each tagged fish in 3D using the same triangulation principals of GPS systems.


Among the tagged fish, there were nine immobile tags, of which seven (or 14% of the total number) were P. argenteus and two (15%) were P. maculatus. These tags were removed from the analyses under the assumption that they represented mortalities or tag loss.

The remainder of the tagged fish left the tailrace before the end of the study period. More than 75% left the tailrace in less than one week, although the average of residence time was longer for P. maculatus (10.79 days) than P. argenteus (2.81 days).

The 66 MW Tres Marias hydroelectric plant, where this study was conducted, is on the Sao Francisco River in southeastern Brazil.
The 66 MW Tres Marias hydroelectric plant, where this study was conducted, is on the Sao Francisco River in southeastern Brazil.

However, 35 individuals (19 P. maculatus and 16 P. argenteus) returned, and most returned more than twice — some up to 38 times. There was no statistically significant difference in the rate of return between the two species. However, there was a significant negative correlation between the number of visits and individual size for P. maculatus, but no such correlation between these two variables for P. argenteus. This result can explain the predominance of smaller individuals of P. maculatus (mean length of 22 cm) among fishes entrained within turbines at Tres Marias Dam, according to CEMIG data.

The average tailrace residency duration was significantly longer for P. maculatus at about 29 hours than P. argenteus, which stayed in the tailrace about six hours. However, there was no correlation between tailrace residency duration and the size of individuals for either species.

Within a day, patterns of movement varied between the two species. P. maculatus were more active during nocturnal periods, whereas P. argenteus’ activity was greater during the daytime. But, analysis of the ratio of inward to outward movements in the tailrace indicated a prevalence of outward movements from the tailrace during the first 14 days of the study. After that the ratio was 1:1.

Lessons learned

Telemetry technology has done much to further the understanding of how dams affect fish migratory patterns. The major threat to fish below hydropower facilities in Brazil is associated with high mortality due to fish entering and becoming entrained within turbine draft tubes.

Although knowledge of fish movement patterns in the tailrace can be an essential step to determine strategies to minimize this impact, applications of acoustic telemetry have proven challenging, with high turbulence, high flows, entrained air and electrical interference proving problematic. Despite the challenges, the acoustic tags and tracking techniques used within the Tres Marias tailrace delivered clear and useful information with a fidelity as fine as 20 cm in 3D position resolution.

The experiment was a feasibility test to see whether the researchers would be able to effectively track fish in the tailrace of a dam. The research team was more than satisfied with the outcome of this feasibility study.

This overhead view shows the placement of the hydrophones used to capture data from the HTI acoustic tags.
This overhead view shows the placement of the hydrophones used to capture data from the HTI acoustic tags.

Due to funding limitations, the study was only conducted for 81 days. This study cost about US$300,000 and was funding by CEMIG. The results of this project have been useful for hydro project owners to define the best scenario of turbine operation with less risk to fish, since the fish mortality in turbines can occur in combination with fish attraction to the tailrace.

Although this is the first study applying this technology in the tailrace of Brazilian hydropower, these results can encourage the use of telemetry techniques to provide the necessary data to allow asset managers to take actions related to hydropower facility operations.

Through continual progress in the development of acoustic tag tracking technology, one will be able to track tailrace fish in more extreme situations and define operational patterns to minimize harm to migratory fish in dam tailraces.

Fabio Suzuki is a fisheries biologist and consultant at Pisces Consultoria e Serviços Ambientais in Brazil. Paulo Santos Pompeu is a professor at Federal University of Lavras and served as advisor for the project.

Previous articleEMEC Supports MHK Technologies on Pathway to Commercialization
Next articleA Hydro Girl at Heart

No posts to display